GIACIATION OF LOWER COPPER RIVER 455 cised.1 From this difference in level, it is thought that the preglacial drainage through the Hanagita Watergap went northward. The watergap is thought of as widened out to the U-shape by the ice which streamed southward from the Copper River basin. The glacial erosion of the watergap produced a secondary valley within it, but this glacial valley nearly everywhere coincides with Wood Canyon and, therefore, is recognizable at only one point, where about half a mile of it is preserved. This is where the Railway leaves the west bank of the river, just north of Mile Post 124, and follows this abandoned valley (c in PL CLXXVI, B, and Fig. 71) which is 80 or 90 feet above the present level of the river. When the glacier was melting away this glacial valley was apparently occupied by a tongue of stagnant ice long enough to prevent the main stream from flowing here, for part of the Wood Canyon channel was established just to the east That glacial waters occupied this channel at one stage, however, is shown by outwash gravels 90 feet above the present Copper River. Wood Canyon (PL CLXXVI) was afterward incised to its present depth and is chiefly a postglacial gorge. In it the Copper River is constricted in a channel 400 to 550 feet wide, in contrast with 1300 feet just below the canyon and over half a mile just north of it. The origin of Wood Canyon is clearly involved with the glacial period rather than orographic movements. The great terraces of outwash gravels in the Copper and Chitina valleys are graded up to an elevation that makes it impossible to consider their formation in relation to anything but streams graded up to a temporary baselevel at the height of the top of the Wood Canyon walls, or the approximate height of the abandoned valley followed by the railway. This establishes Wood Canyon as post-glacial, and suggests the control of the downcutting of the rock floor of the canyon over the system of terraces and gorges which have been cut in the glacial gravels of the whole Copper River basin. How long ago Wood Canyon was eroded cannot be stated, nor is it likely that down-cutting has entirely ceased. It seems certain, however, that it has taken several thousand » See PI. I., Bull. 874, U. S. Geol. Survey. 1909. 400. ZOOO BOO FIG. 71. COPPEB EIVBB IN WOOD CANTON.